Back to the main page.
Bug 1768 - win32bit: nan* mex files missing
| Status | CLOSED FIXED |
| Reported | 2012-10-04 11:45:00 +0200 |
| Modified | 2014-03-12 12:21:39 +0100 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Windows |
| Importance: | P3 blocker |
| Assigned to: | Jörn M. Horschig |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Jörn M. Horschig - 2012-10-04 11:45:21 +0200
and the corresponding .m cannot be executed cause it's a script, not a function (though an empty one) I copied the old nanmean.mexw32 back to svn. gotta see where it needs to be autosynced
Boris Reuderink - 2012-10-04 12:34:10 +0200
Presumably this means that you have access to a Win32 computer. Would you be able to recompile nan{sum, mean, numel, var_base} and commit those? That would resolve this issue.
Jörn M. Horschig - 2012-10-04 12:47:22 +0200
sure, no problem. do you know the locations where the functions need to be autosynced to?
Boris Reuderink - 2012-10-04 14:56:41 +0200
Yes. The same locations as for the Windows 64 bit binaries: $ find . -name "nanmean.mexw64" ./connectivity/private/nanmean.mexw64 ./contrib/spike/private/nanmean.mexw64 ./external/dmlt/external/utils/nanmean.mexw64 ./fileio/private/nanmean.mexw64 ./private/nanmean.mexw64 ./src/nanmean.mexw64 ./statfun/private/nanmean.mexw64 ./utilities/private/nanmean.mexw64 But if that is a hassle I can do it for you — I started the trouble after all :).
Jörn M. Horschig - 2012-10-05 11:01:22 +0200
hehe, it's no hassle, just a matter of a few minutes ;) svn ci connectivity/private/ contrib/spike/private/ fileio/private/ private/ src/ statfun/private/ utilities/private/ -m "enhancement-#1768- re-added nan*.mexw32" Rev. 6698 svn propset autosync true utilities/private/nan*.mexw32 (etc...) svn ci connectivity/private/ contrib/spike/private/ fileio/private/ private/ src/ statfun/private/ utilities/private/ -m "enhancement-#1768- autosynced nan*.mexw32" Rev. 6699 didn't hit the 6700th revision, damn it
Boris Reuderink - 2012-10-05 11:44:18 +0200
It seems that there are still two-versions of the nan* mex files for Win32 in the repo: $ find . -name "nanmean.mexw32" | xargs sha1sum | sort 5b59463614c5bb99018b88af3ab7e2af83da7046 ./contrib/spike/private/nanmean.mexw32 5b59463614c5bb99018b88af3ab7e2af83da7046 ./fileio/private/nanmean.mexw32 5b59463614c5bb99018b88af3ab7e2af83da7046 ./src/nanmean.mexw32 5b59463614c5bb99018b88af3ab7e2af83da7046 ./statfun/private/nanmean.mexw32 5b59463614c5bb99018b88af3ab7e2af83da7046 ./utilities/private/nanmean.mexw32 c6614f0f99368d7e6974ed450900b5af3d3c1354 ./connectivity/private/nanmean.mexw32 c6614f0f99368d7e6974ed450900b5af3d3c1354 ./external/dmlt/external/utils/nanmean.mexw32 c6614f0f99368d7e6974ed450900b5af3d3c1354 ./private/nanmean.mexw32 Note that the latter three files have different checksums. Probably we should not change the external one?
Jörn M. Horschig - 2012-10-05 12:44:39 +0200
right, I didn't touch the external folder (also, I dont have access to it) I don't know how it could end up with different checksums... didn't remex or did anything else with the file http://code.google.com/p/fieldtrip/source/detail?r=6675 http://code.google.com/p/fieldtrip/source/detail?r=6697 http://code.google.com/p/fieldtrip/source/detail?r=6698 http://code.google.com/p/fieldtrip/source/detail?r=6699
Boris Reuderink - 2012-10-05 12:55:43 +0200
Okay, that is very odd. Can you confirm at least that you have different checksums as well? I think my local SVN repo is clean, but to be sure...
Jörn M. Horschig - 2012-10-05 13:12:28 +0200
yep, same for me
Jörn M. Horschig - 2012-10-24 10:17:01 +0200
Boris, any idea what to do with this bug? is it solved?
Robert Oostenveld - 2012-12-06 21:27:45 +0100
I am going over the code for the SVN revision IDs and the copyrights, and noticed that some nanxxx functions have a matlab implementation (as fallback when the mex is not present) and some don't. I suggest to make a fallback implementation for all, which is also to serve as reference implementation.